Utilize or Die: The Supply-Side Gap in Marketplace Models
Low utilization in marketplaces creates a plethora of follow-on problems. How can investors spot them, and how can founders fix them? We asked a few experts within the BAN network.
Consumer-oriented marketplace-based models are probably the predominant type of companies that end up pitching to us. If we had to put a percentage, it might be as high as 60%. That makes sense - the barrier to entry is relatively low, and there are tangible examples in both Bangladesh and around the world of successful ecommerce startups that have gone onto scale and raise tons of money. In an early stage ecosystem like ours, it may make sense to replicate models from other markets, especially for inexperienced founders. And marketplaces can unlock value, when helping to reduce friction present in still developing economies like Bangladesh. This can include information asymmetry and a mismatch between suppliers and demand. The information asymmetry point is appropriate in an economy where so many are entering the middle and affluent class and consuming digitally for the first time. They are discovering how and what to consume, and building habits and preferences. For these consumers, there is value to be unlocked in many verticals through access to information and greater choice, including food, education (e.g. connecting teachers & content to students where they are), niche products (e.g. modest fashion, female hygiene products that aren’t readily available), healthcare (e.g. doctors, information service providers and medication), life-changing big ticket items (e.g. appliances, smart phones, cars, homes), leisure services (e.g. entertainment content, hotel accommodations and plane tickets), etc. Additionally, in disaggregated markets with millions of small, informal businesses, be it mom-and-pop shops, street side pharmacies or even armies of students and stay-at-home moms running facebook commerce pages, marketplaces add a level of filtration through ratings and selection as well as trust through central customer support.
But having listened to a lot of these pitches, quite a few can be summarized along these lines:
We need capital to help grow our revenues
We need to invest in marketing to grow our user base, including potential national expansion, almost always including Chittagong and Sylhet, the 2nd and 3rd largest cities beyond Dhaka
At the same time, we will invest in growing and diversifying our pool of qualified suppliers who, because we’ve trained them, will be dependent on us and therefore will be amenable to our terms including commission rates and payment collection protocols
The growing national user base, along with a supply base, will allow us to convert more of the users into sales
As sales increase, so will our lifetime value (LTV) as we find ways to cross-sell more products/services to our users, especially higher margin ones compared to the original, low-margin products/services we used to hook the customers, as well as squeezing higher commissions from ever-dependent and -dependable suppliers
There are a lot of assumptions within those steps. The one we will focus today is on the supply side, notably utilization.
According to a16z, utilization is the match rate or “the rate at which buyers can find sellers, and vice versa. How to define this metric depends on the unique business.” A lack of utilization, or low match rates, results in the phenomenon of “multi-tenanting,” whereby suppliers will go from one platform to another in search of incremental revenue potential, with very little loyalty. Think of the motorbike rider having multiple ride-hailing apps on, waiting for a match, or restaurants relying on multiple delivery services, or employers listing jobs on multiple portals. In Bangladesh, the online platforms are not only competing with each other for suppliers’ availability (whether it is time in the case of services or inventory in the case of product), but also with the traditional, offline and often informal means of customer acquisition.
For example, there are many service marketplaces connecting informal providers to homes and businesses looking for everyday services like plumbing and cleaning. When marketplaces are unable to provide a sufficient and regular stream of high value work, the providers aren’t staying idle - at various times throughout the day they are going out and getting their own customers by roaming neighborhoods, cutting referral deals with building security or sitting in their shops in neighborhood bazaars with relevant inventory and waiting for customers to come to them for installations and repairs. As a result, they may not be available when a customer orders a service through the online marketplace, which cuts down on the utility of the said marketplace to the consumer, meaning they are less likely to reorder, resulting in high churn and low average lifetime value.
Geography is another factor. Being unable to connect suppliers with customers close enough to their neighborhood may result in long traveling times for the chosen supplier and wait times for customers. For low value work, suppliers aren’t willing to nor can afford to travel too far, and if the marketplace cannot give them jobs close enough, they will not be available when the marketplace calls because they’ll be busy doing work found on their own. Low fulfilment, once again, creates high customer churn.
This fulfilment issue can also happen in product marketplaces. One marketplace we came across, which focuses on custom products and crafts, was having issues with its artisans listing their products not just on their site exclusively but others. There is no guarantee when a customer selects a product for purchase that it is available.
Low utilization of suppliers, when combined with a payment model where the suppliers will collect from customers at the point of service or delivery, and the marketplace has to collect on the back-end from suppliers, is a recipe for disaster. When suppliers control the payment flow and they do not depend on the marketplace for the bulk of their revenues, they will be less incentivized to give up the commission immediately. This also creates room for disintermediation, where the supplier might try to cut out the marketplace and keep all or most of the proceeds, often colluding with the customer by offering them a discount from the marketplace’s listed price, at the expense of the marketplace commission.
So far, we’ve seen that low utilization creates other problems such as multi-tenanting, low fulfilment rates, high churn, disintermediation, payment collection issues and quality issues in fulfilment, not to mention being unable to dictate or negotiate on the commission/take rate. So what is a marketplace company to do? We tend to find two natural reactions. One is to try to burn money on marketing and try to acquire a large enough user base to dictate terms to suppliers for access to those customers. The other is to burn money trying to create a large enough supplier base to be able to churn through low-performing, unreliable suppliers. Both are expensive and may not be the best use of investor money, if the fundamental issues of low utilization and match rates are not solved. All it means is that companies are scaling their problems, because customer and supplier churn will remain high and the businesses will lose more money as they grow.
So what can a founder do? We asked a few marketplace experts within the BAN network.
Go-to-Market: Start with a Wedge, with a Focus on Supplier Acquisition and Fulfillment
There’s a tendency among founders to try to offer a large array of SKUs and products/services right away, in the hopes of enticing a wide range of customers and growing their wallet share.It’s the digital equivalent of a crowded, messy bazaar stall. Supplier acquisition and retention suffers in this process of expanding the offerings too much, too soon.
Kishwar Hashemee, founder of Kludio, an integrated food delivery platform in Bangladesh that combines cloud kitchens, delivery logistics, order interface and marketing for both its own brands and individual restaurants, advises the opposite: “Be the best in class at matching hyper niche suppliers to hyper niche customers. This can help with match rates because the customer acquisition team is optimizing marketing budget to search and convert a specific type of customer, while the merchant acquisition team can focus on identifying and onboarding specific types of merchants. In the early days, this will allow the startup to keep their teams lean and focussed,” while making the most of limited resources. Kishwar was previously VP of Rides at Pathao, a superapp connecting millions in Bangladesh to ride hailing, food, medicine and other services.
Blake Hirt, startup advisor and previously head of operations at Uber Works, agrees: “User growth should be mostly organic if it's a great product [in consumer marketplaces]. And customer LTV is typically fairly low. Both Uber and Airbnb focus much more on supply acquisition, which makes the product better, which organically spurs more demand.” A related and underrated component here is supply-side customer service, which according to Blake is “typically quite poor with marketplaces, due to low investment and low margins. Investing in that early on can be a differentiator and show suppliers that you have their back, earning more trust and loyalty.”
Once you’ve built a wedge, don’t stop there: “Build tools around the wedge: Build the best in class tools for helping the merchants from your niches do MORE on your platform than any other platform. This can help mitigate multi-tenanting on the supply side, and perhaps even lead to informal exclusivity” without the need to either strong arm or entice them through expensive carrots such as pre-payments. Kludio, for example, is building a “kitchen-as-a-service (KaaS)” platform for its merchants to have real-time reporting of their business, including inventory, as well as connecting to logistics providers, among other benefits. On the other hand, “Some marketplaces are providing career growth and skill building opportunities for suppliers. It's not always possible but can be a way to garner loyalty and retention,” according to Blake.
Another word of advice from Kishwar when building products and tools for suppliers: Isolate what’s important. “When planning to build value added solutions to increase stickiness of suppliers, isolate which piece of the full stack matters the most to the suppliers. For example: suppliers might care about revenue 10X more than automated reports and analytics, and 3X more about warehouse support than, once again, analytics. In this case, you should first focus on revenue generation as that is the single most important driver of stickiness,” and put all resources into delivering that consistently for suppliers.
Focusing on a particular niche and supply side fulfilment also has implications for branding. While there is a tendency to confuse branding as simply for consumers only, it matters for both consumers and suppliers. “Position yourself as the best in the world for your [chosen] niche. Etsy has successfully positioned themselves as the sales channel for creative products. Strong brand positioning reduces customer/supplier acquisition costs, and can also help with multi tenanting on the demand side,” according to Kishwar.
As the wedge is built, the focus should be on methodical expansion, into adjacent geographies, markets and categories, all the while making sure supply is keeping up.
Control the Quality and Quantity of Supply, Early On and Potentially Longer
It might be more expensive, but marketplaces might think about integrating suppliers onto the payroll in the beginning in order to ensure reliability and high quality service, rather than relying on a purely distributed model. Pathao is an interesting case. According to Ahmed Fahad, Co-founder and SVP of Product at Pathao, when Pathao started its ride hailing platform in 2016, “we bought new motorcycles and also used existing ones from our courier business. We hired the first ~100 drivers on an hourly basis.” Fast forward a couple of years, Pathao’s ride-hailing platform grew to have millions of users and 100,000’s of freelance riders.
Why did they do that? According to Fahad: “We wanted to ensure reliability, sure. But safety was a bigger concern for first-time customers. We wanted to control the quality of supply to ensure the first 1,000 rides are amazing. Helmets, branded t-shirts, courteous behavior - it all mattered! Also note, motorbike sharing didn't exist as a mode of transport in Bangladesh at the time. Starting off with freelancers may have been inhibitory” and turned off customers from the medium all together through sub-par and unreliable service, even before the market got a chance to take off and consumer habits have been built.
This wasn’t just true for Pathao: “In the early days of Uber, we would simply pay drivers to sit online. They'd get paid whether orders come in or not. Bootstrapping supply is typically needed in the early days to solve the chicken/egg problem, especially when the marketplace is geographically constrained and when the supplier can't multitask while idle” or else there would be danger of them being unavailable when orders did come through, according to Blake.
But does it make sense to have this integrated model for longer? Zantrik, a marketplace connecting individual and fleet owners of vehicles with repair services, recently shifted from a distributed model of independent garages to having its own network through a franchise model. Why did it do that? Shubho Al-Farooque, CEO of Zantrik, elaborates:
“We identified that the core problem was not connecting vehicle owners with technicians, rather, ensuring quality and great experience. The aggregator model was not serving our purpose, as customers were dropping and their satisfaction rate was low as the work quality from garages was not good enough [and Zantrik had limited leverage]. In the Zantrik Digital Garage (franchise model), we are onboarding some carefully vetted garages and upgrading their equipment, training their manpower, establishing our digital systems and applying an efficient process into these garages through training and supervision. These garages will have Zantrik branding and work with Zantrik on a revenue share model across both online and offline orders. The goal is to generate more business for the garages with the same resources by improving transparency and efficiency, without compromising quality, and at the same time ensuring revenue and growth for us. The goal is to onboard 100’s of such garages across the country in the next two years, almost like “OYO rooms for garages.”
Going back to the point about value added services for the merchants, what benefits are they offering the franchised garages and their owners? The major benefit is traceability: Whenever any vehicle enters and exits from the garage, the vehicle owner, garage owner and Zantrik is notified in real time with the vehicle information (number plate and vehicle picture), through a computer vision-guided system that reads number plates and attaches a unique identifier. A “Digital Job Card” is created to trace the workflow of the maintenance process, with service and spare parts information. This gives the customer a piece of mind about their vehicle, the owner a way to remotely track the productivity of their employees and reduce potential leakage and Zantrik with visibility into the overall operations of the garage, as well as its payment flow.
These are just a few examples out of many potential tactics. Here are a few recommended readings, which might be worth their own contextualization for Bangladesh in the future:
Sarah Tavel: Hierarchy of Marketplaces —How to build a marketplace that “wins”. Focus (Level 1), Tip (Level 2), Dominate (Level 3)
Bill Gurley: All Markets Are Not Created Equal: 10 Factors To Consider When Evaluating Digital Marketplaces
Andrew Chen: Required reading for marketplace startups: The 20 best essays
Many thanks to Kishwar Hashemee, Blake Hirt, Ahmed Fahad and Shubho Al-Farooque for generously giving their time and insights.